Being a high performance computing system, a server is usually utilized as a node to store or process data on a network. Comparing to regular personal computers, servers are more powerful in data processing, more stable, reliable, expandable, manageable and secure. Coupled with recent quantum leaps in telecommunication technologies and ever increasing information traffic, servers have become a commodity in corporations and institutions to provide data download, upload and e-mail services. Consequently, improving processing power and reliability so that it can offer a more stable and user-friendly information managing environment is a crucial issue to the quality of servers. Manufacturers continuously put serious effort in resolving malfunctions in servers which may cause interruption of data acquisition or network service. The ultimate goal is to develop new technologies to tackle problems and increase stability.
For example, some servers have Intelligent Platform Management Interface (IPMI) installed to monitor system health and manage the serve system. The IPMI consists of a main controller called a Baseboard Management Controller (BMC). The BMC connects to sensors of devices to be monitored (e.g. CPU for its temperature and voltage or CPU fan for its rotation speed) via a standard system interface such as I2C/IPMB (Intelligent Platform Management Bus/Bridge), Serial/Modem interface or LAN interface, and collects sensed measurements from said sensors. When serious anomaly is determined by said BMC basing on said sensed measurements of monitored devices, BMC may perform corrective actions to allow the server to provide non-stop service, or it may halt the service to prevent more serious mistakes from occurring.
Said BMC comprises firmware for determining whether serious anomaly has happened to a device in said server. Said firmware resides in a storage unit (e.g. flash memory) in the format of an image file. An update of said firmware of said BMC is usually carried out by a boot loader program in a boot sector of said storage unit. Said boot loader runs under a Disk Operating System (DOS) which provides only a simple user interface for said firmware update, therefore a user encounters considerable amount of inconvenience when performing an update of an image file of firmware.
Furthermore, to effectively manage the chassis (or enclosure) and the switch of a server so that said server can provide reliable services and resources, a Roadrunner & Chassis Management System (RCMS) might be installed. An RCMS uses file system and is stored as multiple image files in a storage unit (e.g. hard disk drive). If a user were to update an RCMS via a file system, not only will it be extremely complicated, once the RCMS update fails, there will not be a means to restore back the previous image file.
From the above-mentioned, image file updates are necessary in achieving the goal of providing a more powerful, more stable, more reliable, more expandable, more manageable and more secure serve system. However, said update procedure is highly inconvenience and in the case of an unsuccessful update, the original workable image file is not automatically recovered. Thus, a method for updating an image file which resolve said drawbacks will greatly benefit both the manufacturers and users.